Coccidioides immitis will be used to ask if pathogenesis is related to genotypic variation, if variation in antigenic gene products produced by a pathogen avoids host immunity, if evolutionary selection is occurring in the antigens produced by the pathogen, and if isolates of different virulence can be compared to identify genes necessary to cause disease? Genetic variation related to fungal pathogenesis is blazingly obvious from species to species, e.g., the virulence of C. immitis compared to its close benign relative Uncinocarpus reesii, but it can be more cryptic within a species. Historical studies of C. immitis have known differences in clinical presentation, disease in mice, and abundant morphological variation, but there has been little connection between phenotype and genotype. We are using nucleic acid sequence variation to genetically characterize clinical isolates of C. immitis. To find polymorphic loci, anonymous regions of the genome are amplified from all isolates, and are screened for variation electrophoretically. The variation is confirmed by nucleotide sequencing, and loci and alleles are as precise and unambiguous as possible. Work in progress shows that sequence variation is abundant and easily found within and among populations of C. immitis; genetic variation correlates with geographic location and isolates from one site are variable and show a range of relatedness. To support the program of vaccine production, the C. immitis isolates of the range of genotypes will be tested in mice to determine variation in pathogenicity. To rule out antigens that are not universally expressed, the set of C. immitis isolates will be tested for transcription of antigen genes. To ensure that protection is achieved throughout the range of pathogen diversity, immunized mice will be challenged by a representative set of isolates. Analysis of pathogen evolution and of the genetic basis of pathogenicity will follow from the tests in mice and from the study of antigen production. Selection should be strong on pathogen gene products recognized by the host. We can compare the nucleotide variation in genes for antigenic proteins within C. immitis and between C. immitis and U. reesii to detect this selection. If we find strong variation in pathogenicity (aided by new environmental isolates from soil and small mammals). we can use differences in the genome or its transcripts to search for the determinants of pathogenicity.